Advancing High-Resolution Imaging of Virus Assemblies in Liquid and Ice

Liza-Anastasia DiCecco; Samantha Berry; G. M. Jonaid; Maria J. Solares; Liam Kaylor; Jennifer L. Gray; Carol Bator; William J. Dearnaley; Michael Spilman; Madeline J. Dressel-Dukes; Kathryn Grandfield; Sarah M. McDonald Esstman; Deborah F. Kelly

doi:10.3791/63856

JoVE Journal > Biochemistry

Biochimie

Advancing High-Resolution Imaging of Virus Assemblies in Liquid and Ice

Published: July 20, 2022

doi:

10.3791/63856

Liza-Anastasia DiCecco*^1,2, Samantha Berry*¹, G. M. Jonaid*^1,3, Maria J. Solares^1,4, Liam Kaylor^1,4, Jennifer L. Gray⁵, Carol Bator⁶, William J. Dearnaley¹, Michael Spilman⁷, Madeline J. Dressel-Dukes⁸, Kathryn Grandfield², Sarah M. McDonald Esstman⁹, Deborah F. Kelly^1,5,6

¹Department of Biomedical Engineering,Pennsylvania State University, ²Department of Materials Science and Engineering,McMaster University, ³Bioinformatics and Genomics Graduate Program, Huck Institutes of the Life Sciences,Pennsylvania State University, ⁴Molecular, Cellular, and Integrative Biosciences Graduate Program, Huck Institutes of the Life Sciences,Pennsylvania State University, ⁵Materials Research Institute,Pennsylvania State University, ⁶Huck Institutes of the Life Sciences,Pennsylvania State University, ⁷Applications team,Direct Electron, ⁸Application Scientist,Protochips, Inc., ⁹Department of Biology,Wake Forest University

Summary

Here protocols are described to prepare virus assemblies suitable for liquid-EM and cryo-EM analysis at the nanoscale using transmission electron microscopy.

Abstract

Interest in liquid-electron microscopy (liquid-EM) has skyrocketed in recent years as scientists can now observe real-time processes at the nanoscale. It is extremely desirable to pair high-resolution cryo-EM information with dynamic observations as many events occur at rapid timescales – in the millisecond range or faster. Improved knowledge of flexible structures can also assist in the design of novel reagents to combat emerging pathogens, such as SARS-CoV-2. More importantly, viewing biological materials in a fluid environment provides a unique glimpse of their performance in the human body. Presented here are newly developed methods to investigate the nanoscale properties of virus assemblies in liquid and vitreous ice. To accomplish this goal, well-defined samples were used as model systems. Side-by-side comparisons of sample preparation methods and representative structural information are presented. Sub-nanometer features are shown for structures resolved in the range of ~3.5-Å-10 Å. Other recent results that support this complementary framework include dynamic insights of vaccine candidates and antibody-based therapies imaged in liquid. Overall, these correlative applications advance our ability to visualize molecular dynamics, providing a unique context for their use in human health and disease.

Introduction

Biomedical research improves our understanding of human health and disease through the development of new technologies. High-resolution imaging is transforming our view of the nanoworld – permitting us to study cells and molecules in exquisite detail¹^,²^,³^,⁴^,⁵. Static information of dynamic components such as soft polymers, protein assemblies, or human viruses reveals only a limited snapshot of their complex narrative. To better understand how molecular entities operate, their structure and function must be jointly investigated.

Recent advances in the production of materials such as atomically thin graphene or silicon-based microchips provide new opportunities for real-time structure-function analysis using transmission electron microscopes (TEMs). These materials can create hermetically sealed chambers for live EM imaging⁶^,⁷^,⁸^,⁹^,¹⁰^,¹¹. The new field of liquid-EM, the room temperature correlate to cryo-EM, provides unprecedented views of hard or soft materials in solution, allowing scientists to simultaneously study the structure and dynamics of their specimen. Liquid-EM applications include real-time recordings of therapeutic nanoparticles interacting with cancer stem cells as well as changes in the molecular intricacies of viral pathogens¹²^,¹³^,¹⁴.

Just as methodological advances spurred the resolution revolution in the cryo-EM field, new techniques and methods are needed to extend the use of liquid-EM as a high-throughput tool for the scientific community. The overall goal of the methods presented here is to streamline liquid-EM specimen preparation protocols. The rationale behind the developed techniques is to employ new microchip designs and autoloader devices, suitable for both liquid- and cryo-EM data collection (Figure 1)⁷^,¹⁴^,¹⁵^,¹⁶^,¹⁷. The assemblies are mechanically sealed using standard grid clips for automated instruments, such as the Krios, which can accommodate multiple samples per session or a F200C TEM (Figure 2). This methodology expands the use of high-resolution imaging beyond standard cryo-EM applications demonstrating broader purposes for real-time materials analysis.

In the current video article, protocols are presented for preparing virus assemblies in liquid with and without commercially available specimen holders. Using the specialized specimen holder for liquid-EM, thin liquid specimens can provide structural information comparable to cryo-EM samples, as well as dynamic insights of the specimens. Also demonstrated are methods for preparing liquid specimens using autoloader tools for high-throughput routines. The major advantage over other techniques is that automated specimen production allows the user to quickly assess their samples for optimum thickness and electron dosage prior to data collection. This screening technique quickly identifies ideal areas for real-time recordings in liquid or ice¹²^,¹⁴^,¹⁸^,¹⁹. For purposes of 3D structure determination, liquid-EM may complement the long-established cryo-EM methods implemented in cryo-EM. Readers employing conventional TEM or cryo-EM technologies may consider using liquid-EM workflows to provide new, dynamic observations of their samples in a manner that complements their current strategies.

Virus samples used in this protocol include purified adeno-associated virus subtype 3 (AAV) obtained as a gift and cultured under standard conditions¹². Also used were non-infectious SARS CoV-2 sub-viral assemblies derived from the serum of COVID-19 patients¹² and obtained from a commercial source. Finally, purified simian rotavirus (SA11 strain) double-layered particles (DLPs) were obtained from the laboratory of Dr. Sarah M. McDonald Esstman at Wake Forest University and cultured using standard conditions ⁶^,¹⁷. Software packages described here are freely available and the links have been provided in the Table of Materials section.

Protocol

1. Loading the specimen holder for liquid-EM

Clean the silicon nitride (SiN) microchips by incubating each chip in 150 mL of acetone for 2 min followed by incubation in 150 mL of methanol for 2 min. Allow chips to dry in laminar airflow.
Plasma clean the dried chips using a glow-discharge instrument operating under standard conditions of 30 W, 15 mA for 45 s using Argon gas.
Load a dry base microchip into the tip of the specimen holder. Add ~0.2 µL of sample (0.2-1 mg/mL of virus assemblies in 50 mM HEPES, pH 7.5; 150 mM NaCl; 10 mM MgCl₂; 10 mM CaCl₂) to the base chip. Following a 1-2 min incubation step, place the top chip on the wet base chip containing the sample.
Clamp the assembly together to form a hermetically sealed enclosure, held in place mechanically by three brass screws. Upon sealing the assembly, pump the tip to 10^-6 Torr using a turbo-pumped dry pumping station. The holder is now ready to be inserted into the TEM.
NOTE: The select holder has no cooling capabilities and is not used for cryo-EM.

2. Production of microchip sandwich assemblies

NOTE: Different SiN or silicon dioxide (SiO) microchips can be used directly from the shipped gel packs. Carbon-coated gold grids may also be used directly as supplied.

Plasma clean the microchips and carbon grids using a glow discharge instrument operating under standard conditions of 30 W, 15 mA for 45 s using Argon gas.
Add ~2 µL of sample contained in 50 mM HEPES, pH 7.5; 150 mM NaCl; 10 mM MgCl₂; 10 mM CaCl₂ to a glow-discharged microchip placed on a gel pack. Remove excess solution (~50%) using a filter paper or a pipette. Following a 1-2 min incubation step, add the glow-discharged carbon grid to the wet microchip containing the sample.
Clamp the assembly together using a single-tilt specimen holder or autoloader grid clips at room temperature to form a hermetically sealed enclosure. For the auto-loader clamps, place the sandwich assembly on the bottom C-clip, place the top clip on top of the assembly and use the standard clamping tool to seal the assembly together.
NOTE: The clamping procedure performed here uses the same steps as for cryo-EM grid clamping, but at room temperature. Clamped samples may be stored for 2 months or longer prior to imaging while maintaining liquid in the enclosure. No leak check is used with a room temperature specimen holder or the autoloader.
The specimen is now ready to be inserted into the TEM. Examine samples placed in autoloaders under cryo conditions or at room temperature.

3. Imaging specimens using a transmission electron microscope

Liquid-EM imaging
1. Load the specimen holder into the TEM equipped with a field-emission gun (FEG) and operating at 200 kV.
2. Turn on the gun and adjust the eucentric height of the microscope stage with respect to the specimen by using the wobbler function, tilting the sample from -15° to +15° in the column. This procedure adjusts the stage in the Z-direction to accommodate the sample thickness. And helps ensure an accurate magnification during image recording.
3. Record images as long-framed movies using the in situ package integrated with a direct detector having a pixel spacing of 6 µm (Video 1 and Video 2). Individual images may also be recorded using the serial data collection software package²⁰, implementing automated imaging routines. Acquire images under low-dose conditions at magnifications ranging from 28,000x-92,000x and 40 frames per second.
4. Adjust exposure times (0.25-1 s) to minimize beam damage to the specimen. Use a defocus range of -1-4 µm at the specified magnification. If a thick solution is encountered, use higher defocus values or select a different region of interest.
5. Ensure the solution is present in the samples throughout the imaging session by focusing the electron beam on a sacrificial area not used for data collection, until bubbles are formed (Supplementary Figure 1).
Cryo-EM imaging
1. Load the clipped EM grids or microchip sandwiches into the TEM equipped with a FEG and operating at 300 kV. Turn on the gun and adjust the eucentric height of the microscope stage, using a similar procedure described for the liquid-TEM above (step 3.1.2).
2. Record individual images using the single particle analysis system integrated within the microscope system while implementing automated imaging routines. Record images under low-dose conditions using the single particle analysis direct electron detector having a pixel spacing of 14 µm at a magnification of 59,000x and 40 frames per second.
3. Use a defocus range of 1-4 µm at the specified magnification. If thick layers of vitreous ice are encountered, use higher defocus values, or select a different region for data collection.

4. Data analysis and 3 D structure comparisons

Analysis of adeno-associated virus (AAV) in liquid and vitreous ice
1. Process movies for AAV particles in liquid and ice using RELION-3.08 program²¹ or other image processing software²². Perform motion correction using MotionCor2 v1.2.3.
2. Once corrected, extract particles using the auto-picking tool in the program software package. Typical box sizes are 330 pixels for liquid specimens and 350 pixels for ice specimens.
3. Calculate initial reconstructions using C1 symmetry using the program's 3D initial model routine and/or the ab-initio model options in the data processing software package. In the program, use a regularization parameter of T = 4 and a pixel size of 1.01 Å for liquid specimens and 1.13 Å for ice specimens.
4. Use a mask value of 300 Å throughout the refinement procedures. Perform refinement protocols in the data processing software using I1 symmetry to obtain multiple EM maps. Expected structural resolution may be in the range of 4 Å or better. Use particle equivalency implementing icosahedral symmetry at 16,800 for liquid and 15,240 for ice¹²^,¹⁴.
  NOTE: Resolution estimates are based on the gold-standard Fourier shell correlation (FSC) criteria.
5. Mask EM maps at ~250 Å and examine results using a molecular structure analysis software package²³^,²⁴. In Figure 3, slices are shown in increments of ~5 nm.
6. Extract capsid protein (VP1) subunits from the EM maps for comparison. Quantify dynamic changes among EM structures based on the particle diameter measurements (Supplementary Figure 2), and then visualize using the Morph map function in the software.
Analysis of SARS-CoV-2 sub-viral assemblies in liquid
1. Process movies using the RELION-3.08 program. Correct for image drift and beam-induced motion using MotionCor2 v1.2.3. Correct for the contrast transfer function (CTF) of the microscope.
2. Use the auto-picking tool in the program to select viral assemblies with a box size of 800 pixels. For computational efficiency, extracted particles can be rescaled to 256 pixels. Obtain an initial model using C1 symmetry in the program with a regularization parameter of T = 2 and 1.66 Å pixel size.
3. Perform 3D refinement in the program to obtain an EM map, an example is shown at ~8.25 Å according to standard FSC criteria (Figure 4). Visualize EM maps using the molecular structure analysis software with slices incremented at 25 nm as demonstrated in Figure 4.
Analysis of rotavirus double-layered particles (DLPs) in vitreous ice
1. Process movies for rotavirus DLPs using other image processing software. Use MotionCor2 v1.2.3 to correct for drift in the images. Use Patch CTF estimation to correct for lens effects on the images.
2. Extract particles using the auto-picking tool in the software with a box size of 950 pixels. Down-sample the box size as needed for computing purposes.
3. Calculate initial models using ab-initio options and C1 symmetry. Use refinement parameters including a pixel size of 1.47 Å and a mask value of 800 Å.
4. Perform additional refinement routines while imposing I1 symmetry. Example results include a 10.15 Å EM map, according to the gold-standard FSC criteria. Total particles used were 2050, which equates to 123,000 protomer units due to the icosahedral symmetry operator (Figure 5).
5. Mask final maps at ~750 Å and visualize results in the molecular structure analysis software package. Example slices through the EM map are shown in Figure 5 in increments of ~10 nm.

Representative Results

A liquid-TEM operating at 200 kV was used for all liquid-EM imaging experiments and a cryo-TEM operating at 300 kV was used for all cryo-EM data collection. Representative images and structures of multiple viruses are presented to demonstrate the utility of the methods across various test subjects. These include recombinant adeno-associated virus subtype 3 (AAV), SARS-CoV-2 sub-viral assemblies derived from the patient serum, and simian rotavirus double-layered particles (DLPs), SA11 strain. First, comparisons are demonstrated for the same AAV sample (1 mg/mL) imaged in liquid buffer solution (50 mM HEPES, pH 7.5; 150 mM NaCl; 10 mM MgCl₂; 10 mM CaCl₂) and in vitreous ice (Figure 3A–C; Supplementary Figure 2). The liquid-EM samples were prepared using SiN-based microchips and the specialized specimen holder. The cryo-EM samples were prepared using standard holey carbon grids and vitrified using a state-of-the-art specimen preparation unit into liquid ethane. The resulting virus structures had similar overall diameters of ~25 nm. A comparison of individual VP1 capsid protomers also showed consistent features in both liquid and ice EM maps (Figure 3D). One difference between the liquid-EM and cryo-EM data was that additional dynamic structures were observed in liquid that were not present in the cryo-EM results (Figure 3E)¹².

Next, examined were deactivated SARS-CoV-2 samples (0.25 mg/mL) prepared in buffer solution (50 mM HEPES, pH 7.5; 150 mM NaCl; 10 mM MgCl₂; 10 mM CaCl₂) using a new microchip sandwich technique⁷^,¹⁴. The new technique employs SiN-based microchips along with carbon-coated gold grids. In this preparation, the liquid sample was sandwiched between the two substrates (Figure 2A–C). Sandwiched specimens were clamped with a single-tilt specimen holder or autoloader C-clips commonly used in cryo-EM sample preparation. Samples can be viewed immediately in the TEM or stored at 4°C for 2 months or beyond, depending upon the stability of the biological specimen. The appearance of bubbling in the liquid samples ensures the presence of the liquid layer. Liquid thickness can be measured using EF-TEM protocols as described¹². SARS-CoV-2 sub-viral assemblies in liquid appeared to have high visible contrast with respect to the surrounding liquid (Figure 4A,B). Some particles displayed visible spikes on the virus surface while some particles are lacking spikes or are sparsely decorated with them (Supplementary Figure S3). Class averages and slices through the 8.25 Å structure showed internal features within the particles comprising protein subunits and the viral RNA genome (Figure 4C). Although some particles appeared to contain C2 symmetry (upon testing C2 symmetry, the particle equivalency was 2,674), the asymmetric structure did not differ much in comparison to symmetrical structure.

Finally, the analysis included rotavirus DLPs (3 mg/mL; 50 mM HEPES, pH 7.5; 150 mM NaCl; 10 mM MgCl₂; 10 mM CaCl₂) in vitreous ice by manually flash-freezing microchip sandwich preparations into liquid nitrogen. The same sandwich technique used to produce liquid-EM samples was employed for frozen-hydrated specimens. Sandwiches were sealed with autoloader grid clips and examined using the cryo-TEM under standard conditions. Low-magnification views of the frozen specimens showed little to no cubic or hexagonal ice contamination and regions of densely packed DLPs were observed throughout the viewing windows (Figure 5A,B). Images were collected using automated routines implemented in the EPU package. Class averages and slices through the 10.15 Å map revealed stable features consistent with viral protein components and the double-stranded RNA genome (Figure 5D,E). The contrast difference between the DLPs and the ice background was not as visibly strong in the cryo-EM images compared with the liquid-EM samples. While the cause of this interesting effect is still being determined, it is worth noting the difference as liquid-EM may offer future advantages for the imaging community.

Figure 1: Techniques used for high-resolution imaging of viruses in liquid and in ice. The two workflows highlight different liquid-EM sample preparation methods. The left panel shows a schematic representation of a liquid-EM specimen holder. The SiN base microchip includes an array (500 µm x 100 µm) of integrated microwells (10 µm x 10 µm) that are ~150 nm in-depth with a ~30 nm thick membrane. The right panel presents a schematic of the microchip sandwich technique, which can be used for both liquid-EM and cryo-EM research. The sandwich assemblies use a SiN microchip paired with a carbon-coated gold TEM grid. Across-section of the assembly indicates multiple imaging windows ranging from 250 µm x 250 µm to 50 µm x 50 µm in size with membrane thicknesses of 10 nm or 5 nm, respectively. The carbon support film is ~5 nm thick. Please click here to view a larger version of this figure.

Figure 2: The microchip sandwich technique for liquid-EM and cryo-EM. (A) The microchip sandwich assembly uses a SiN microchip paired with a carbon-coated gold TEM grid. A glow-discharged microchip is placed on a gel pack and virus samples are added to the microchip. After a brief incubation period, excess solution is removed, and the sandwich is sealed with the TEM grid. (B) The microchip sandwich is sealed in a clipping device at room temperature and can be loaded directly into a single-tilt TEM holder or a TEM autoloader system. (C) Cross-section drawings of the microchip sandwich assembly highlight the dimensions of the microchips and carbon layer. Please click here to view a larger version of this figure.

Figure 3: Comparison of liquid-EM and cryo-EM structures of AAV. (A) Structure of AAV in solution (3.22 Å resolution) with colored radial densities showing 5 nm slices through the map. Scale bar is 5 nm. Imaging metrics are for data acquisition using the direct electron detector. (B) Structure of AAV imaged in ice (3.37 Å resolution) with colored radial densities represent 5 nm slices through the structure. Scale bar is 5 nm. Imaging metrics are for data acquisition using the direct electron detector. (C) A region of interest shows 5 s and 20 s time points along with Fourier transforms calculated at different time points. Left side shows CTF estimates, right side shows the experimental data. (D) Rotational views of the AAV VP1 subunit extracted from the liquid and ice structures. The segments were interpreted using the crystal structure (PDB code 3KIC, A chain²⁵). Scale bar is 10 Å. (E) Dynamic values in the liquid structures generated using the morph map function in the molecular structure analysis software and described in step 4.1.6-4.2.3. From left to right, averaged structures from multiple virus assemblies show conformational changes with a corresponding ~5% diameter change, measured using EM data. RMSD values in voxels indicate changes according to the color scale. The figure has been modified from¹². Please click here to view a larger version of this figure.

Figure 4: SARS-CoV-2 sub-viral assemblies prepared in liquid using the microchip sandwich technique. (A) Image of SARS-CoV-2 sub-viral assemblies isolated from serum fractions from COVID-19 patients. Imaging metrics are for data acquisition using the direct electron detector. White bubbles in the top-right corner of the image are a visual indicator that liquid is present in the sample. Scale bar is 100 nm. (B) Calculated Fourier transform of the image shows high-resolution information in reciprocal space with little to no drift in the corresponding image. Class averages show unique features in the viral assemblies contained in the solution. (C) An EM reconstruction of the sub-viral assemblies is shown with colored radial densities at 5 nm slices through the map. Scale bar is 25 nm. Please click here to view a larger version of this figure.

Figure 5: Rotavirus DLPs prepared in liquid using the microchip sandwich technique. (A,B) Low magnification screening steps using the analysis software. Limited ice crystals were observed, and window membranes were thin and clear, simplifying area selection for data acquisition. Scale bar in (A) is 5 µm and in (B) is 500 nm. (C) Movies were acquired using the direct electron detector in counting mode according to the indicated imaging metrics. Scale bar is 50 nm. (D) Fourier transform information indicates high-resolution information present in the data and class averages show strong features in the icosahedral lattice. Class averages show features in the virus assemblies free of artifacts and consistent with other structural observations⁶^,¹⁵^,¹⁷. (E) An EM reconstruction of the DLPs (10.15 Å resolution) shown with colored radial densities at 10 nm slices through the map. Scale bar is 15 nm. Please click here to view a larger version of this figure.

Supplementary Figure 1: Confirming the presence of liquid in EM samples. Low-magnificationimages of AAV particles in solution acquired under low dose conditions (<10 e^– Å^-2 s^-1) lack bubbles. Using a focused beam of >100 e^– Å^-2 s^-1, liquid samples showed bubble formation (black arrows). Scale bar is 500 nm. The figure has been modified from¹². Please click here to download this File.

Supplementary Figure 2: Measuring changes in AAV dimensions in liquid. (A) Contour traces of virus particles were acquired over a 20 s time frame. Scale bar is 25 nm. (B,C) Table and graphical representations of particle diameter measurements at 1, 5, 10, and 20 seconds. The overall change in particle diameters during the 20 s recording was 0.28% at a dose of ~1e^– Å^-2 s^-1. (D,E) Signal-to-noise values determined at varying time point during movie acquisition. The figure has been modified from¹². Please click here to download this File.

Supplementary Figure 3: High-magnification image of patient-derived SARS-CoV-2 sub-complexes. Heterogenous viral assemblies showed the presence of spike proteins (red arrows) on the surface of some particles. Other particles were lacking in surface spikes as labeled. Scale bar is 100 nm. Please click here to download this File.

Video 1: Real-time imaging of AAV in liquid. Side-by-side comparison of a real-time recording of AAV in liquid (left) and colored contour traces of the particles diffusing in solution (right). Please click here to download this Video.

Video 2: Visualization of image tracing in liquid. Side-by-side comparisons of a colored and low-pass filtered movie of AAV diffusing in solution (left) colored contour traces of the particles diffusing in solution (right). Please click here to download this Video.

Discussion

New opportunities are presented to streamline current liquid-EM workflows by using new automated tools and technologies adapted from the cryo-EM field. Applications involving the new microchip sandwich technique are significant with respect to other methods because they enable high-resolution imaging analysis in liquid or vitreous ice. One of the most critical steps in the protocol is producing specimens with the ideal liquid thickness to visualize exquisite details at the nanoscale level. Ideal regions of interest are identified at lower magnifications by screening the entire sample prior to implementing high-throughput routines for automated data collection. Should the identified regions in liquid or ice specimens contain beam-damage artifacts or appear too thick to identify individual particles that are visually crisp, they should be excluded from data collection. Limitations in high-resolution data acquisition include beam-induced movement and Brownian motion in the liquid samples. Cryo-EM methods aim to minimize motion in biological samples; however, motion correction methods are computationally robust at mitigating these resolution-limiting factors. If the liquid samples present long-range drift and thermal instability, thinner samples may be prepared by removing excess solution during sample preparation steps to decrease the thickness of the liquid layers.

Molecular crowding among virus particles may result when samples are used at high concentrations. If there are multiple regions with overlapping particles in a manner that limits robust data collection, the input sample should be reduced in concentration prior to sample preparation. Having a suitable sample concentration is a crucial step in the protocol. Images that contain an excess of particles will complicate downstream image processing procedures. For purified virus particles, an adequate concentration range is between 0.3-3 mg/mL, depending on the dimensions and molecular weight of the sample. Larger viruses (~100 nm diameter or greater) typically require higher concentrations than do smaller viruses (~25 nm diameter or lesser) for the success of the technique. Another factor to consider is the level at which the particles of interest adhere to the glow-discharged microchips. If the virus sample does not adhere well to the surface, a greater concentration may be needed to adequately prepare specimens for liquid-EM or cryo-EM. Alternatively, the sample can instead be applied to the carbon-coated gold grid with the microchip serving as the lid of the enclosure. Should these procedures fail to recruit sufficient particles for data collection procedures, affinity capture methods may present a viable option²⁶^,²⁷^,²⁸.

The use of certain additives such as detergents, glycerol, polyethylene glycols, and high levels of sucrose or glucose should be minimized or avoided for liquid-EM imaging. These reagents may introduce artifacts in the images as well as lead to excessive bubbling, hydrolysis products, and free radical formation due to beam damage. Such effects lead to muted features in the imaged particles and ultimately limit structural resolution. One way to remove these reagents if they are used in biochemical preparations is through extensive dialysis into a buffer solution lacking the additives. These reagents will need to be tested and used on a case-by-case basis. Moreover, once sufficient specimens have been produced using the microchip sandwich technique, they may be imaged right away or stored at 4 °C until ready to examine. Storage times depend upon sample stability.

Overall, using liquid-EM in combination with cryo-EM permits researchers to examine biological systems with complementary imaging tools. The newly developed microchip sandwich technique yields consistent samples for TEM imaging in liquid or ice. The technique also provides a simple means for researchers to prepare and view specimens without the need for a commercial sample holder or vitrification system. In combination with automated imaging protocols, large amounts of data in the order of thousands of images per session may be acquired per specimen. The pioneering work of Parsons and colleagues²⁹^,³⁰^,³¹^,³²^,³³ laid the foundational groundwork for producing biological specimens in liquid enclosures. The protocols presented here describe how state-of-the-art tools can now provide an exciting means to visualize biological macromolecules through new eyes. Future applications that are expected to result from mastering these techniques, when paired with high-performance computing, are real-time mechanistic insights for understanding structure-function relationships in 3D. The liquid-EM field may serve to elevate research on novel viruses that pose a threat to human health, perhaps even contributing to pandemic preparedness measures. In summary, the use of these protocols should permit scientists and engineers to better study dynamic processes at atomic detail, across a large variety of samples encompassing life sciences, medicine, and materials research.

Divulgations

The authors have nothing to disclose.

Acknowledgements

The authors acknowledge Dr. Luk H. Vandenberghe (Harvard Medical School, Department of Ophthalmology) for providing purified AAV-3. This work was supported by the National Institutes of Health and the National Cancer Institute (R01CA193578, R01CA227261, R01CA219700 to D.F.K.).

Materials

Acetone	Fisher Scientific	A11-1	1 Liter
Autoloader clipping tool	ThermoFisher Scientific	N/A	Also SubAngstrom supplier
Autoloader grid clips	ThermoFisher Scientific	N/A	top and bottom clips
Carbon-coated gold EM grids	Electron Microcopy Sciences	CF400-AU-50	400-mesh, 5-nm thickness
COVID-19 patient serum	RayBiotech	CoV-Pos-S-500	500 microliters of PCR+ serum
Methanol	Fisher Scientific	A412-1	1 Liter
Microwell-integrad microchips	Protochips, Inc.	EPB-42A1-10	10×10-mm window arrays
TEMWindows microchips	Simpore Inc.	SN100-A10Q33B	9 large windows, 10-nn thick
TEMWindows microchips	Simpore, Inc.	SN100-A05Q33A	9 small windows, 5-nm thick
Top microchips	Protochips, Inc.	EPT-50W	500 mm x 100 mm window
Whatman #1 filter paper	Whatman	1001 090	100 pieces, 90 mm
Equipment
DirectView direct electron detector	Direct Electron		6-micron pixel spacing
Falcon 3 EC direct electron detector	ThermoFisher Scientific		14-micron pixel spacing
Gatan 655 Dry pump station	Gatan, Inc.		Pump holder tip to 10^-6range
Mark IV Vitrobot	ThermoFisher Scientific		state-of-the-art specimen preparation unit
PELCO easiGlow, glow discharge unit	Ted Pella, Inc.		Negative polarity mode
Poseidon Select specimen holder	Protochips, Inc.		FEI compatible;specimen holder
Talos F200C TEM	ThermoFisher Scientific		200 kV; Liquid-TEM
Titan Krios G3	ThermoFisher Scientific		300 kV; Cryo-TEM
Freely available software	Website link		Comments (optional)
cryoSPARC	https://cryosparc.com/		other image processing software
CTFFIND4	https://grigoriefflab.umassmed.edu/ctffind4		CTF finding program
MotionCorr2	https://emcore.ucsf.edu/ucsf-software
RELION	https://www3.mrc-lmb.cam.ac.uk/relion/index.php?title=Main_Page
SerialEM	https://bio3d.colorado.edu/SerialEM/
UCSF Chimera	https://www.cgl.ucsf.edu/chimera/		molecular structure analysis software package

References

Deng, W., et al. Assembly, structure, function and regulation of type III secretion systems. Nature Reviews Microbiology. 15 (6), 323-337 (2017).
Oikonomou, C. M., Chang, Y. -. W., Jensen, G. J. A new view into prokaryotic cell biology from electron cryotomography. Nature Reviews Microbiology. 14 (4), 205-220 (2016).
Murata, K., Wolf, M. Cryo-electron microscopy for structural analysis of dynamic biological macromolecules. Biochimica et Biophysica Acta. General Subjects. 1862 (2), 324-334 (2018).
DiMaio, F., et al. Atomic accuracy models from 4.5 Å cryo-electron microscopy data with density-guided iterative local refinement. Nature Methods. 12 (4), 361-365 (2015).
Frank, J., et al. A model of protein synthesis based on cryo-electron microscopy of the E. coli ribosome. Nature. 376 (6539), 441-444 (1995).
Dukes, M. J., Gilmore, B. L., Tanner, J. R., McDonald, S. M., Kelly, D. F. In situ TEM of biological assemblies in liquid. Journal of Visualized Experiments. (82), e50936 (2013).
Dearnaley, W. J., et al. Liquid-cell electron tomography of biological systems. Nano Letters. 19 (10), 6734-6741 (2019).
Park, J., et al. Direct observation of wet biological samples by graphene liquid cell transmission electron microscopy. Nano Letters. 15 (7), 4737-4744 (2015).
Chen, Q., et al. 3D Motion of DNA-Au nanoconjugates in graphene liquid cell electron microscopy. Nano Letters. 13 (9), 4556-4561 (2013).
Yuk, J. M., et al. High-resolution EM of colloidal nanocrystal growth using graphene liquid cells. Science. 336 (6077), 61-64 (2012).
Wang, X., Yang, J., Andrei, C. M., Soleymani, L., Grandfield, K. Biomineralization of calcium phosphate revealed by in situ liquid-phase electron microscopy. Communications Chemistry. 1, 80 (2018).
Jonaid, G., et al. High-resolution imaging of human viruses in liquid droplets. Advanced Materials. 33 (37), 2103221 (2021).
Pohlmann, E. S., et al. Real-time visualization of nanoparticles interacting with glioblastoma stem cells. Nano Letters. 15 (4), 2329-2335 (2015).
Jonaid, G. M., et al. Automated tools to advance high-resolution imaging in liquid. Microscopy and Microanalysis. , 1-10 (2022).
Varano, A. C., et al. Customizable cryo-EM chips improve 3D analysis of macromolecules. Microscopy and Microanalysis. 25, 1310-1311 (2019).
Alden, N. A., et al. Cryo-EM-on-a-chip: Custom-designed substrates for the 3D analysis of macromolecules. Small. 15 (21), 1900918 (2019).
Tanner, J. R., et al. Cryo-SiN – An alternative substrate to visualize active viral assemblies. Journal of Analytical and Molecular Techniques. , (2016).
Solares, M. J., et al. Microchip-based structure determination of disease-relevant p53. Analytical Chemistry. 92 (23), 15558-15564 (2020).
Casasanta, M. A., et al. Microchip-based structure determination of low-molecular weight proteins using cryo-electron microscopy. Nanoscale. 13 (15), 7285-7293 (2021).
Mastronarde, D. N. Advanced data acquisition from electron microscopes with SerialEM. Microscopy and Microanalysis. 24, 864-865 (2018).
Scheres, S. H. W. RELION: Implementation of a Bayesian approach to cryo-EM structure determination. Journal of Structural Biology. 180 (3), 519-530 (2012).
Punjani, A., Rubinstein, J. L., Fleet, D. J., Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nature Methods. 14 (3), 290-296 (2017).
Pettersen, E. F., et al. UCSF Chimera—A visualization system for exploratory research and analysis. Journal of Computational Chemistry. 25 (13), 1605-1612 (2004).
Goddard, T. D., Huang, C. C., Ferrin, T. E. Visualizing density maps with UCSF Chimera. Journal of Structural Biology. 157 (1), 281-287 (2007).
Lerch, T. F., Xie, Q., Chapman, M. S. The structure of adeno-associated virus serotype 3B (AAV-3B): Insights into receptor binding and immune evasion. Virology. 403 (1), 26-36 (2010).
Sharma, G., et al. Affinity grid-based cryo-EM of PKC binding to RACK1 on the ribosome. Journal of Structural Biology. 181 (2), 190-194 (2013).
Kiss, G., et al. Capturing enveloped viruses on affinity grids for downstream cryo-electron microscopy applications. Microscopy and Microanalysis. 20 (1), 164-174 (2014).
Degen, K., Dukes, M., Tanner, J. R., Kelly, D. F. The development of affinity capture devices—a nanoscale purification platform for biological in situ transmission electron microscopy. Rsc Advances. 2 (6), 2408-2412 (2012).
Hui, S. W., Parsons, D. F. Electron diffraction of wet biological membranes. Science. 184 (4132), 77-78 (1974).
Hui, S. W., Parsons, D. F., Cowden, M. Electron diffraction of wet phospholipid bilayers. Proceedings of the National Academy of Sciences of the United States of America. 71 (12), 5068-5072 (1974).
Parsons, D. F., Matricardi, V. R., Moretz, R. C., Turner, J. N. Electron microscopy and diffraction of wet unstained and unfixed biological object. Advances in Biological and Medical Physics. 15, 161-270 (1974).
Parsons, D. F. Structure of wet specimens in electron microscopy. Improved environmental chambers make it possible to examine wet specimens easily. Science. 186 (4162), 407-414 (1974).
Matricardi, V. R., Moretz, R. C., Parsons, D. F. Electron diffraction of wet proteins: Catalase. Science. 177 (4045), 268-270 (1972).

Play Video

PDF

DOI

DOWNLOAD MATERIALS LIST

Citer Cet Article

DiCecco, L., Berry, S., Jonaid, G. M., Solares, M. J., Kaylor, L., Gray, J. L., Bator, C., Dearnaley, W. J., Spilman, M., Dressel-Dukes, M. J., Grandfield, K., McDonald Esstman, S. M., Kelly, D. F. Advancing High-Resolution Imaging of Virus Assemblies in Liquid and Ice. J. Vis. Exp. (185), e63856, doi:10.3791/63856 (2022).